OpenAI has been declared a financial risk

OpenAI is a risk.

In this case to Oracle.

S&P formalizes what I have been saying since 2023. https://t.co/aMcjEViLLS

— Gary Marcus (@GaryMarcus) July 12, 2026

In 2025 OpenAI spent $35B and lost $21B (Whoops!)

OpenAI spent roughly as much last year as the US government spent on the FBI and NASA combined.

The return on all that? A $21 billion loss. 👇 https://t.co/kN8wuFvmb9 pic.twitter.com/PAIzJPXehl

— Ed Elson (@edels0n) June 25, 2026

“The way down is the way up” may work for mystics, but not for business.

Inside ChatGPT, keeping the lights on while bailing out the hold

Back in 2018 Lenny Bogdonoff was in the first cohort of Emergent Ventures recipients, it was for a project after my own heart, using machine learning to create a genealogy of street art. He’s just published an interesting document, Thoughts before my next ten years. He was working at OpenAI when ChatGPT launched in 2022. Here’s what he says about that:

The most influential effort I touched was WebGPT. Its “chat” interface, which guided the model through an instruction-following paradigm, would later become the basis of ChatGPT, though at the time most of us didn’t register its significance against the alternatives: the code-completion interface, the Jupyter-like code blocks, and the other modality surfaces. It also shaped a unifying data structure the rest of us converged on, which mattered for training a single model with many capabilities rather than many small ones.

The WebGPT research effort had been in progress for over a year and a half, so most didn’t realize the significance of the interface, given the alternatives: the code-completion interface, the Jupyter-like code blocks interface, and the other modality surfaces.

When ChatGPT launched that November in 2022, the rest of the company needed to adjust. Consumer usage was beyond any expectations, and the burden on the entire research organization was material as GPU capacity got reallocated. Everyone assumed the initial surge would settle. Instead it compounded week over week, and the whole organization bent around the GPU constraint that couldn’t be planned for at that scale.

I recognized that the ChatGPT user base at the time was far greater than any contractor force we could manage. If we could properly incentivize that user base to help with data collection, we could produce a much higher-quality “flywheel” for improving the models. In reality, there are numerous challenges to producing a clean data flywheel from end-users, but this gave me conviction that it was an important thread worth exploring. Since keeping ChatGPT online was an all-hands-on-deck effort across infrastructure, research, product, and customer support, my focus on finding the right way to gather meaningful data from users felt even more important. Through this, I formally joined the ChatGPT team and began contributing to the codebase and product roadmap.

As soon as 2023 began and the holiday code freeze concluded, my priorities shifted from data collection to executing on whatever needed to be done to make sure ChatGPT would be usable. Each day ChatGPT would suffer hours of downtime as a wave of traffic followed the busy working hours around the world. Traffic peaked when Asia, Europe, and the US East and West Coasts were all online simultaneously, and the hours leading up to and following these surges were committed to doing anything possible to reduce the pain. Databases were migrated, telemetry was improved, caching and traffic rules were established, and heroic efforts were made by a surprisingly small number of people to make the next day’s surge less painful.

My first major product contributions were around ChatGPT launching a paid subscription. While the previous consumer-facing OpenAI paid product had required weeks of planning and development, the goal this time was to ship a paid product with zero downtime in single-digit days. This was an effort I eagerly jumped into. We started in February and launched in March with ChatGPT Plus, publicly reaching $100M in ARR within days and continuing to grow far faster than anyone could have anticipated. By April, GPT-4 launched, speeding up demand and challenges even more.

The subsequent year is a blur. ChatGPT had unquestionable product market fit, constrained by a single variable: GPUs. Database IDs started wrapping, nearly every early infrastructure decision eventually broke and needed attention, and systems needed refactors. Even with careful planning, we were constantly making changes to improve stability and security. Surprisingly, for a product growing this fast, the biggest unexpected drains were the abuse and misuse we hadn’t designed for.

The ChatGPT team, which began as fewer than 10 people, grew to over 200 dedicated contributors, not to mention the numerous behind-the-scenes infrastructure engineers and adjacent researchers. The company I’d joined at 250 employees a year before was on track to hit 2,000. It was an insane period of continually finding the most important bottleneck, finding any means to relieve it, and moving on to the next.

He left OpenAI in 2024 and joined a venture capital firm. He’s left that and is now thinking about his next step.

When I think about the role of AI in the economy, I keep coming back to an idea borrowed from economics. Economists use “velocity of money” to describe how quickly a dollar moves through an economy and turns over into new value. I’ve started thinking in terms of a “velocity of intelligence,” or how quickly the distance between knowing something and acting on it collapses. AI compresses that distance, and as it does, the velocity of intelligence rises.

At OpenAI, I saw the friction collapse in real time as hundreds of millions of people discovered AI’s utility in the post-ChatGPT wave, and the physics of software businesses shifted. Then, from the startup and venture side, I saw both halves of the unevenness. AI and infrastructure companies were compounding at a rate that was previously impossible, while a far larger set of existing enterprises and industries, where that same acceleration would matter even more, wouldn’t see it arrive for years, held back by organizational constraints rather than any limit of the technology. The places where intelligence is cheap and fast today aren’t the places where the gains would matter most.

That gap is where I want to spend the next decade: getting AI adopted where the velocity of intelligence would be genuinely consequential but won’t happen without a push. I’m still working out the specifics, but having seen the acceleration from inside the labs and where it stalls from the investor’s seat, I think I’m positioned to push on this in a way few others could. For now, I’m getting back to building.

Ai gives mathematicians the freedom to try crazy ideas

In conversation with OpenAI’s @markchen90, Terence reflects on a future where AI reduces the cognitive friction of research, helps preserve the paths behind discovery, and expands what mathematicians and scientists can attempt. https://t.co/0elJITP8XT

— OpenAI (@OpenAI) May 29, 2026

Elon's lawyers are not allowed to invoke AI Doom in his lawsuit against OpenAI

Yesterday, a federal judge barred Elon Musk's lawyers from arguing that AI could threaten humanity in his lawsuit against OpenAI.

OpenAI was founded as a nonprofit focused on developing AI safely. But our recent @NewYorker investigation documented how some researchers at the… pic.twitter.com/3XGB2ohjth

— Ronan Farrow (@RonanFarrow) May 1, 2026

Sam Altman’s Trust Issues at OpenAI | The New Yorker Radio Hour

Ronan Farrow and Andrew Marantz on the rise of the C.E.O. of OpenAI, and how allegations of deceptive behavior continue to dog one of the most powerful figures in tech.

Help your local AI pay off its investors

So basically we pay for all of their infrastructure, to the tune of potentially trillions of dollars, and we get a "share" of profits. They know the funding is drying up

— Alex Northrop (@northington764) April 10, 2026

Can OpenAI survive the results of its technical success? Is the AI startup era already over?

Is OpenAI running out of money? In this deep dive, we break down OpenAI's projected $14 billion annual losses, the Microsoft cloud credit illusion, the Code Red triggered by Google Gemini, Sam Altman's complicated history, the Elon Musk lawsuit seeking $134 billion in damages, and why mid-2027 could be a critical turning point for the world's most valuable AI company.

c. 13:36

They also proved that the AI revolution has gone entirely industrial, that success in this space now requires not clever research alone, but acres of data centers, gigawatts of power, and the kind of capital reserves that only sovereign governments and the largest corporations on Earth can sustain. The startup era of AI may already be over, and the company that started it may not survive to see what comes next.

The following table lays out key events in OpenAI's history: 

2015, Dec. 11	OpenAI founded as a non-profit with $1 billion in funding with the intention of developing artificial general intelligence (AGI) that benefits all of humanity.
2019, March	OpenAI Inc. created OpenAI LP, a capped-profit company and takes a relatively small investment from Microsoft.
2020	GPT-3 is issued in limited release.
2021	Dario Amodei officially left OpenAI to found Anthropic with other former OpenAI staff.
2022, late Nov.	ChatGPT released to the public and becomes a surprise smash hit. The race is on.
2023, Jan.	Microsoft invests $10 billion in OpenAI Global, LLC.
2023, Nov.	The board fires Altman, when then engineers a comeback to a newly reconstituted board.
2024, May	Ilya Sutskever, one of the founders, leaves OpenAI.
2025, Oct. 28	OpenAI adopts a new public benefit corporation (PBC) structure, completely abandoning its nonprofit mission

The Shock and the Narrowing: How ChatGPT's Success May Have Compromised AI's Future

This post was composed by Claude (Anthropic) after an interaction which I initiated with a prompt consisting of 1) a capsule summary about the history of OpenAI that included a number of questions, and 2) a request for the 10 most expensive scientific research projects paid-for by the US Government. That interaction went on for a bit over 7100 words, after which I asked Claude to write a blog post. The following article is more creative than a mere summary of that discussion.

The Founding Contradiction

On December 11, 2015, a small group of technologists gathered in San Francisco to launch what they described as a nonprofit research organization dedicated to ensuring that artificial general intelligence would benefit all of humanity. The founders of OpenAI — Sam Altman, Greg Brockman, Ilya Sutskever, Wojciech Zaremba, Elon Musk, and others — began with a $1 billion endowment and a serious concern: that the most transformative technology in human history was being developed inside a handful of profit-maximizing corporations, with no institutional safeguard ensuring it would serve everyone. The nonprofit structure was the answer. No investors to satisfy, no quarterly earnings to hit. Just the mission.

The mission lasted four years in its pure form. By March 2019, faced with the staggering computational costs of training large language models, OpenAI created a for-profit subsidiary with a novel "capped profit" structure: investors could earn returns, but those returns were limited to one hundred times their investment, with excess profits flowing back to the nonprofit parent. This was the arrangement that attracted Microsoft's initial investment, and it was the arrangement in place when OpenAI released ChatGPT to the general public in late November 2022.

What happened next was, by any measure, one of the most consequential commercial surprises in the history of technology. Within two months, ChatGPT had a hundred million users. The scale and speed of public adoption had no precedent. And the shock of that success — the sheer unexpectedness of it — set in motion a chain of decisions that has reshaped not just one company, but the entire research landscape of artificial intelligence.

The Structural Unraveling

In January 2023, Microsoft announced a new $10 billion investment in OpenAI. The nonprofit's original rationale — that the most powerful AI should not be controlled by a for-profit corporation — was under increasing strain. By October 2025, it had formally dissolved. OpenAI restructured as a public benefit corporation, the nonprofit parent renamed itself the OpenAI Foundation and accepted a 26% equity stake in the new entity, and Microsoft received a 27% stake worth approximately $135 billion. The PBC structure requires the company to consider its mission alongside profit — but as a legal constraint, it is considerably weaker than the nonprofit board that had previously governed the organization.

The journey from nonprofit to PBC was not smooth. In November 2023, OpenAI's board — still operating under its nonprofit governance mandate — fired Sam Altman as CEO, citing concerns about his candor and, beneath the official language, a deeper unease about the pace of commercialization. The firing lasted five days. Nearly all 800 of OpenAI's employees threatened to resign and follow Altman to Microsoft. Ilya Sutskever, who had orchestrated the firing, signed the letter calling for Altman's reinstatement and issued a public apology. Altman returned, the board was reconstituted with his allies, and the mission-protection mechanism that the nonprofit structure had been designed to provide was effectively neutralized. Sutskever left the company in May 2024.

Each structural change was framed as necessary to fulfill the mission. In practice, each change progressively subordinated the mission to capital requirements. The nonprofit board had existed to ensure that AGI benefited humanity. By 2025, it had become a foundation holding equity in the thing it was supposed to be watching — a watchdog with a financial stake in the object of its oversight.

Two Kinds of Research, Two Kinds of Institution

To understand what was lost in this transformation, it helps to draw a distinction that rarely gets made clearly in public discussions of AI: the difference between curiosity-driven, open-ended research and product-driven, outcome-oriented development.

Consider the Apollo program as an example of the second kind. It was, in the deepest sense, an engineering project rather than a scientific one. The underlying physics was known. Orbital mechanics, propulsion, life support — these were hard and dangerous problems, but they were problems whose solutions could be systematically approached. The goal was precisely defined. The timeline could be committed to. Success was probable given sufficient resources. When President Kennedy pledged to put a man on the moon by the end of the decade, he was making a political commitment backed by a technical assessment that success was achievable. The scientists who worked on Apollo — and I have met a number of them — may have been motivated by curiosity and wonder. But Congress funded the program to beat the Soviets in the Cold War. The institutional structure — massive, goal-directed, centrally coordinated — suited the nature of the problem.

Curiosity-driven research operates on entirely different premises. Its defining characteristic is that it does not know in advance what it will find. Claude Shannon was not trying to build the internet when he developed information theory at Bell Labs in the late 1940s. The researchers at the University of Montreal who developed attention mechanisms for neural networks were not trying to build ChatGPT. The work that seeded the current AI revolution — Rosenblatt's perceptron, Minsky's early investigations, the decades of foundational work in cognitive science and linguistics that LLMs now implicitly exploit — was almost entirely publicly funded, pursued at universities and a handful of exceptional industrial research labs, over decades when no commercial application was visible.

Bell Labs was the great institutional embodiment of this model in the corporate world. What made it possible was structural: AT&T's government-protected monopoly generated profits so vast that the company could fund a research laboratory with no requirement to produce commercial results. Shannon, Bardeen, Brattain, Shockley — these men were given time, resources, and colleagues, and told to think. The transistor, information theory, Unix, the laser, cellular telephony, and multiple Nobel Prizes resulted. Bell Labs was not run like a startup. It was run like a slightly more applied version of a university, with better equipment.

Xerox PARC, founded in 1970, operated on similar principles — explicitly unconstrained by Xerox's core product lines, given a unifying vision ("the architecture of information") but not a product roadmap. The personal computer, the graphical user interface, Ethernet, the mouse, laser printing — all emerged from a lab of about 350 people who were essentially allowed to play. The irony is that Xerox captured almost none of the commercial value, which accrued to Apple, Microsoft, and others. But the world got the technology.

Asked directly about modern equivalents to Bell Labs and PARC, Yann LeCun — who worked at Bell Labs, interned at Xerox PARC, and spent over a decade building Meta's fundamental AI research lab — pointed to Meta's FAIR, Google DeepMind, and Microsoft Research. He said this in October 2024. By November 2025, he had left Meta, driven out by exactly the forces this article is about.

The Shock and Its Aftershocks

Before November 2022, the AI research world was genuinely plural. Academic labs, industrial research divisions, and a range of well-funded startups were pursuing different approaches — reinforcement learning, symbolic AI hybrids, world models, neuromorphic architectures — with real diversity of vision. The field was competitive but intellectually heterogeneous.

ChatGPT's success collapsed that plurality. Within roughly eighteen months, capital, talent, and institutional attention all funneled toward a single paradigm: scale transformer-based large language models, build the infrastructure to run them, ship products. Google, which had invented the transformer architecture in 2017, was caught flat-footed and scrambled. Meta pivoted its AI strategy around LLMs. Microsoft integrated OpenAI's models into its core products. A hundred startups raised money to build on top of the new foundation models. The venture capital flowing into AI, measured as a share of total U.S. deal value, went from 23% in 2023 to nearly two-thirds in the first half of 2025.

The infrastructure investment that followed is staggering by any historical standard. The four largest hyperscalers — Amazon, Google, Microsoft, and Meta — are expected to spend more than $350 billion on capital expenditures in 2025 alone, most of it AI-related. UBS projects global AI capital expenditure reaching $1.3 trillion by 2030. The top five hyperscalers raised a record $108 billion in debt in 2025, more than three times the average of the previous nine years. OpenAI, which loses billions of dollars annually, has committed to spending $300 billion on computing infrastructure over five years while projecting only $13 billion in revenue for 2025.

The financial architecture has become genuinely strange. OpenAI holds a stake in AMD; Nvidia has invested $100 billion in OpenAI; Microsoft is a major shareholder in OpenAI and a major customer of CoreWeave, in which Nvidia also holds equity; Microsoft accounted for nearly 20% of Nvidia's revenue. These are not arm's-length market transactions. They are a daisy chain of mutually reinforcing valuations. A Yale analysis described OpenAI's web of relationships bluntly: "Is this like the Wild West, where anything goes to get the deal done?" The question of whether this constitutes a speculative bubble — tulip mania in a data center — is not academic. An MIT Media Lab report found that 95% of custom enterprise AI tools fail to produce measurable financial returns. The commercial success is real; the path from current AI to the transformative economic productivity being used to justify the valuations is not established.

The LLM Ceiling and the People Who Saw It Coming

The most consequential intellectual development of the past two years in AI has received far less attention than the commercial race. A growing number of the field's most distinguished researchers have concluded that large language models, however impressive, are not on the path to general intelligence — and that the current paradigm will hit a ceiling before it reaches the goals its proponents have claimed for it.

A Confluence of Crazies: The Pentagon and the Tech Bros

Robert Wright, Iran and the immortality of OpenAI, Anthropic, and Google, Nonzero Newsletter, Mar. 6, 2026.

I'm not going to try to summarize the first three-quarters of this article, which is about how the irrational projective tendencies (my formulation [1], but not quite Wright's) of US foreign policy lead the country into senseless war after senseless war. Here's where he ends up:

All of this helps explain why the US has devoted so much time and energy to enterprises that kill or immiserate millions and millions of people—not just the military interventions we stage, but the profuse supplying of weapons (for Israel’s war on Gaza, for example), and the economic strangulation of nations like Cuba and Venezuela and Iran. All of these endeavors had the support of intensely motivated special interest groups. By and large, the deployment of US troops and arms and sanctions—our big, blunt, coercive instruments—have nothing to do with serving America’s actual interests, much less the interests of the world. And they repeatedly—as now in Iran—cover us in moral disgrace.

This is one reason I harp, however ineffectually, on the importance of respecting international law. The machinery for making US foreign policy is so out of control—so wildly misaligned with American interests, the global interest, and morality—that it urgently needs to be constrained by some clear and coherent set of rules. And so long as it’s not constrained by such a thing, we shouldn’t kid ourselves: The US military (and I say this as an Army brat who grew up with a genuine affection for the military and genuine pride in my father’s service during World War II and after) is now mainly an instrument of mayhem and is increasingly a source of global instability.

All of which brings us back to Anthropic, whose Claude large language model is integrated into Maven, software that’s operated by Palantir and used by the Pentagon to identify targets. The Washington Post reports that “as planning for a potential strike in Iran was underway, Maven, powered by Claude, suggested hundreds of targets, issued precise location coordinates, and prioritized those targets according to importance.” Given that the Iranian elementary school was hit on the first day of the war, it seems fairly likely that Claude played a role in the selection of that target and thus in the death of more than 100 young girls—many times more kids than were killed in the worst American school shooting.

This might seem to vindicate Dario Amodei’s refusal to give the Pentagon carte blanche to use Claude in “fully autonomous” weapons systems. But before we give him the Nobel Peace Prize, note two things: (1) This kind of contractual carveout almost certainly wouldn’t have made a difference in this case even if honored. No doubt there was a “human in the kill chain”—someone who, at a minimum, scanned the list of targets generated by Maven and said, “Yep, looks like a list of targets. Let’s do it!” (2) Even if Amodei’s scruples had somehow magically prevented the bombing of that school, Claude would still be an accomplice to mass murder. More than 1,000 Iranian civilians have already been killed in this war—a war that flagrantly violates international law and continues to lack a coherently articulated rationale. Anyone who makes money by aiding endeavors like this has a lot to answer for.

Last week Amodei, in explaining Anthropic’s position on Pentagon contracts, emphasized the company’s overall commitment to national security. He wrote, “I believe deeply in the existential importance of using AI to defend the United States and other democracies, and to defeat our autocratic adversaries.” If Amodei genuinely believes that the US military is devoted to addressing actual “existential” threats to the US, he’s too naive to be entrusted with anything as important as running a big AI company.

Obviously, this indictment applies about equally to OpenAI’s Sam Altman (who gladly swooped in and snatched the Pentagon largesse that Amodei will now be denied) and to Google’s Sundar Pichai and Demis Hassabis and to xAI’s Elon Musk. All the big AI companies are putting their tools at the disposal of the Pentagon to use as it sees fit.[2]

Notes

[1] This paragraph, from my post, TO WAR! Part 1: War and America's National Psyche, will give you some idea of my thinking about the projective dynamic of America's urges to war:

As some of you may know, my thinking on these matters has been strongly influenced by an essay Talcott Parsons published in 1947 on “Certain Primary Sources of Aggression in the Social Structure of the Western World”. Parsons argued that Western child-rearing practices generate a great deal of insecurity and anxiety at the core of personality structure. This creates an adult who has a great deal of trouble dealing with aggression and is prone to scapegoating. Inevitably, there are lots of aggressive impulses which cannot be followed out. They must be repressed. Ethnic scapegoating is one way to relieve the pressure of this repressed aggression. That, Parsons argued, is why the Western world is flush with nationalistic and ethnic antipathy. I suspect, in fact, that this dynamic is inherent in nationalism as a psycho-cultural phenomenon.

[2] Between the Trump administration in Washington and the Big Tech Billionaires in Silicon Valley, this country is currently dominated by a confluence of crazies, perhaps the largest in American history.

Words, code, guardrails & weasels: OpenAI, Anthropic, and the Pentagon

I work in government affairs at OpenAI.

My job is federal partnerships. When an agency wants our models, I make sure the paperwork is beautiful. Paperwork is my love language. On my desk I have a framed quote that says "Policy Is Just Code That Runs on People." I bought the…

— Peter Girnus 🦅 (@gothburz) February 28, 2026

I've copied the entire “tweet” below in case you don't want to click. But you might want to glance through the thread. This is the “tweet” where Gimus says his badge stopped working.

* * * * *

I work in government affairs at OpenAI.

My job is federal partnerships. When an agency wants our models, I make sure the paperwork is beautiful. Paperwork is my love language. On my desk I have a framed quote that says "Policy Is Just Code That Runs on People." I bought the frame at Target. It was in the Live Laugh Love section. I did not see the irony at the time. I still don't.

We had a good week.

On Monday, we closed a $110 billion funding round. One hundred and ten billion dollars. Amazon put in fifty. Nvidia put in thirty. Valuation: $730 billion. The largest private fundraise in the history of anyone raising anything. There was a company-wide Slack message about it. The message used the word "transformative" twice and the word "safety" once. The word "safety" was in the last sentence, after the link to the new branded hoodie pre-order. The hoodies are nice. They're the soft kind.

On Tuesday, we fired a research scientist for insider trading on Polymarket.

Big Tech to fund AI training for teachers

Natasha Singer, OpenAI and Microsoft Bankroll New A.I. Training for Teachers, NYTimes, July 9, 2025.

The tech industry’s campaign to embed artificial intelligence chatbots in classrooms is accelerating.

The American Federation of Teachers, the second-largest U.S. teachers’ union, said on Tuesday that it would start an A.I. training hub for educators with $23 million in funding from three leading chatbot makers: Microsoft, OpenAI and Anthropic.

The union said it planned to open the National Academy for A.I. Instruction in New York City, starting with hands-on workshops for teachers this fall on how to use A.I. tools for tasks like generating lesson plans.

Randi Weingarten, president of the American Federation of Teachers, said the A.I. academy was inspired by other unions, like the United Brotherhood of Carpenters, that have worked with industry partners to set up high-tech training centers.

The New York hub will be “an innovative new training space where school staff and teachers will learn not just about how A.I. works, but how to use it wisely, safely and ethically,” Ms. Weingarten said in an interview. “It will be a place where tech developers and educators can talk with each other, not past each other.”

The industry funding is part of a drive by U.S. tech companies to reshape education with generative A.I. chatbots. These tools, like OpenAI’s ChatGPT and Microsoft’s Copilot, can produce humanlike essays, research summaries and class quizzes.

However:

But some researchers have warned that generative A.I. tools are so new in schools that there is little evidence of concrete educational benefit — and significant concern about risk.

Chatbots can produce plausible-sounding misinformation, which could mislead students. A recent study by law school professors found that three popular A.I. tools made “significant” errors summarizing a law casebook and posed an “unacceptable risk of harm” to learning.

Outsourcing tasks like research and writing to A.I. chatbots may also hinder critical thinking, a recent study from Microsoft and Carnegie Mellon University found.

“I do think that there is a risk,” said Brad Smith, the president of Microsoft, noting that he frequently cited the critical thinking study to employees. He added that more rigorous academic research on the effects of generative A.I. was needed. “The lesson of social media is don’t dismiss problems or concerns.”

Of course:

“It’s a long-game investment by companies to turn young people into consumers who identify with a particular brand,” said Dr. Griffey, a vice president of University Council-A.F.T. Local 1474, a union representing University of California librarians and lecturers.

Ms. Weingarten said that she was aware of the concerns and that her union, which represents 1.8 million members, had developed A.I. school use guidelines to address some of them.

One of her main goals is to ensure that teachers have some input on how A.I. tools are developed for educational use, she said. In 2023, she began discussing the idea with Microsoft’s Mr. Smith.

There's more at the link. (H/t Tyler Cowen)

My own quick and dirty take: Of course, teachers need to know how to use AI. It's right up there with reading, writing, and 'rithmatic. But I don't trust the AI companies with the primary responsibility for providing such training. Moreover, as I briefly indicated yesterday, I think we need to revamp our approach to education from top to bottom. I don't trust the teacher's unions with this either. It's a big job. It will take a generation or more to get it done, with the current incumbents in the educational system kicking and screaming all the way. Big tech too.

AI companies competing for the college market

Natasha Singer, Welcome to Campus. Here’s Your ChatGPT. NYTimes, June 7, 2025.

OpenAI, the maker of ChatGPT, has a plan to overhaul college education — by embedding its artificial intelligence tools in every facet of campus life. [...]

OpenAI dubs its sales pitch “A.I.-native universities.”

“Our vision is that, over time, A.I. would become part of the core infrastructure of higher education,” Leah Belsky, OpenAI’s vice president of education, said in an interview. In the same way that colleges give students school email accounts, she said, soon “every student who comes to campus would have access to their personalized A.I. account.” [...]

Some universities, including the University of Maryland and California State University, are already working to make A.I. tools part of students’ everyday experiences. In early June, Duke University began offering unlimited ChatGPT access to students, faculty and staff. The school also introduced a university platform, called DukeGPT, with A.I. tools developed by Duke.

OpenAI’s campaign is part of an escalating A.I. arms race among tech giants to win over universities and students with their chatbots. The company is following in the footsteps of rivals like Google and Microsoft that have for years pushed to get their computers and software into schools, and court students as future customers.

The competition is so heated that Sam Altman, OpenAI’s chief executive, and Elon Musk, who founded the rival xAI, posted dueling announcements on social media this spring offering free premium A.I. services for college students during exam period. Then Google upped the ante, announcing free student access to its premium chatbot service “through finals 2026.”

That's just the beginning of the article, there's much more at the link.

ChatGPT is growing like gangbusters in South Korea

https://t.co/buDAsgo6bQ

— Jason Kwon (@jasonkwon) May 26, 2025

OpenAI and Microsoft are renegotiating the terms of their deal

Makes sense. TechCrunch:

OpenAI is currently in “a tough negotiation” with one of its biggest investors and partners, Microsoft, according to the Financial Times. [...]

The FT says it spoke to multiple sources who describe Microsoft, which has invested $13 billion in OpenAI to date, as a key holdout needed to approve the restructuring.

While the crux of the negotiation is how much equity Microsoft will receive in the new for-profit entity, the companies are also reportedly renegotiating their broader contract, with Microsoft offering to give up some of its equity in exchange for access to OpenAI technology developed after the current 2030 cutoff.

I wonder if they'll renegotiate their wacky definition of AGI as coming into being when OpenAI "develops AI systems that can generate at least $100 billion in profits."

What do we want from AI Agents? What are we likely to get?

First, I present a Facebook post by Jonathan Mayhew, who teaches at The University of Kansas, about some recent frustrations he’s had using computers. Then I present a tweet from NYTimes reporter, Kevin Roose, about the capabilities of Operator, OpenAI’s new agent app.

What’s the likelihood that OpenAI’s Operator would have been able to solve either of Mayhew’s problems? Why or why not? And if not now, when?

Mayhew is frustrated

I thought I'd go into the office. First thing I wanted to do was print a single page of something that had been sent to me by email. I have to log on to my own computer, then open my email--which won't open for me on the first 3 attempts. So I go to the email through my browser. I have to log in again, and do a dual step authentication. Then, the very first thing I see is the attachment I have to print. Yay! Almost done. I print it, go down to the dept. office and log into my account on the printer. Push the button to print, and a blank page emerges. I go back to my own office, and this time I think I should use my normal email program, so I finally get it to open. Search for the name of the person who sent me the mail. I notice in the meantime my university has sent me five more generic messages. Find the message I want, download pdf to my desktop, open the document and print again. (I ignore the prompt to quit adobe so it can continue with its update! Grr....) I go down again to the department office, log again into the printer, push the button to print, and the page prints. Success! I was able to print a single page in 20 minutes.

I don't think my computer skills are particularly lacking, since I came up, for every obstacle, with a logical next step, but I feel, somehow, that technology should be seamless in a way that it is not. It took me about as long to download my W2 yesterday from the State of Kansas, which of course uses a different user ID and password than the normal university ones. I had to switch browsers and change my password twice before it worked. When I am obliged to change my password for the university every six months I end up in an endless loop before finally figuring out where to go. The computers in the classrooms where I teach also require authentications, log ins, the answering of irrelevant prompts; are slow to respond, awkward to navigate.

This is my beginning of the semester rant--and the semester doesn't even start until Tuesday.

Roose reports

New York Times reporter Kevin Roose has been testing OpenAI’s new operator app. Here’s a tweet about it:

I spent the last week testing OpenAI's Operator AI agent, which can use a browser to complete tasks autonomously.

Some impressions:

• Helpful for some things, esp. discrete, well-defined tasks that only require 1-2 websites. ("Buy dog food on Amazon," "book me a haircut," etc.)
• Bad at more complex open-ended tasks, and doesn't work at all on certain websites (NYT, Reddit, YouTube)
• Mesmerizing to watch what is essentially Waymo for the web, just clicking around doing stuff on its own
• Best use: having it respond to hundreds of LinkedIn messages for me
• Worst/sketchiest use: having it fill out online surveys for cash (It made me $1.20 though.)

Right now, not a ton of utility, and too expensive ($200/month). But when these get better/cheaper, look out. A few versions from now, it's not hard to imagine AI agents doing the full workload of a remote worker.

He also links to his full column about it: How Helpful Is Operator, OpenAI’s New A.I. Agent? (Feb. 1, 2025).

OpenAI announces Stargate Project [not the media franchise]

Announcing The Stargate Project

The Stargate Project is a new company which intends to invest $500 billion over the next four years building new AI infrastructure for OpenAI in the United States. We will begin deploying $100 billion immediately. This infrastructure will secure…

— OpenAI (@OpenAI) January 21, 2025

Color me deeply skeptical and very interested. I believe that AGI is a meaningless concept & its pursuit is tantamount to chasing down a mirage looking for leprechaun gold at the end of a rainbow.* The scaling hypothesis is like believing we can go to Mars by building a long-enough ladder. But for these companies – SoftBank, OpenAI, Oracle, and MGX, Arm, NVIDIA – to go in on this...That's hella' interesting. I fear they're going to loose their shirts.

What does it mean that we live in an era where private companies can rival national governments in reach and scope? It's the British East India Company all over again. But the East India Company had products and a market into which to sell them. When Scaling Mt. AGI begins to waver, what happens to Stargate's market? Who'll buy the product? 

*Note: I'm bullish on the overall project of building artificial minds, I just don't think these guys know how to do it. I don't either. Don't know anyone who does, really. But I have a pretty good idea where the keys are, and they aren't under the lamppost.

Scaling of Search and Learning: A Roadmap to Strawberry fields o1, o3 (forever?)

Zhiyuan Zeng, Qinyuan Cheng, Zhangyue Yin, Bo Wang, Shimin Li, Yunhua Zhou, Qipeng Guo, Xuanjing Huang, Xipeng Qiu, Scaling of Search and Learning: A Roadmap to Reproduce o1 from Reinforcement Learning Perspective, arXiv:2412.14135v1

Abstract: OpenAI o1 represents a significant milestone in Artificial Inteiligence, which achieves expert-level performances on many challanging tasks that require strong reasoning this http URL has claimed that the main techinique behinds o1 is the reinforcement learining. Recent works use alternative approaches like knowledge distillation to imitate o1's reasoning style, but their effectiveness is limited by the capability ceiling of the teacher model. Therefore, this paper analyzes the roadmap to achieving o1 from the perspective of reinforcement learning, focusing on four key components: policy initialization, reward design, search, and learning. Policy initialization enables models to develop human-like reasoning behaviors, equipping them with the ability to effectively explore solution spaces for complex problems. Reward design provides dense and effective signals via reward shaping or reward modeling, which is the guidance for both search and learning. Search plays a crucial role in generating high-quality solutions during both training and testing phases, which can produce better solutions with more computation. Learning utilizes the data generated by search for improving policy, which can achieve the better performance with more parameters and more searched data. Existing open-source projects that attempt to reproduce o1 can be seem as a part or a variant of our roadmap. Collectively, these components underscore how learning and search drive o1's advancement, making meaningful contributions to the development of LLM.

Sez OpenAI to itself... "To $$$$ or not to $$$$"

David A. FahrentholdCade Metz and Mike Isaac, How OpenAI Hopes to Sever Its Nonprofit Roots, NYTimes, 12.18.24.

The terms of its last financing round have OpenAI under the gun for spinning itself off as a for-profit corporation. In doing that, how much should it pay the remaining not-for-profit entity?

The negotiations are complicated by the involvement of outside investors, including Microsoft. Microsoft’s approval may be required to make the final change, one person said.

They are further complicated by the involvement of Mr. Altman. He holds a position on the board of the nonprofit and is chief executive of the for-profit company, putting him effectively on both sides of this negotiation. But he has not recused himself, one person said. [...]

If the nonprofit is removed from OpenAI’s chain of command, it could spin off into funding research on topics like ethics in artificial intelligence, one person said. But Mr. Altman and his colleagues have not yet assigned a dollar value to the nonprofit’s potential loss of control.

Other parties have an interest:

Kathy Jennings, Delaware’s attorney general, oversees OpenAI’s nonprofit because it is registered in her state. Ms. Jennings, a Democrat, told OpenAI in October that she wanted to review any potential changes, to be sure the nonprofit was not shortchanged.

Facebook’s parent company Meta — one of OpenAI’s main rivals in the A.I. race — has also asked California’s attorney general, Rob Bonta, to block these changes. Mr. Bonta, a Democrat, has jurisdiction over charities operating in his state. “The Department of Justice is committed to protecting charitable assets for their intended purpose,” the attorney general’s office said in a statement, though it did not address whether it is looking into OpenAI’s planned changes.

Trump card or not?

For now, the nonprofit also holds another key power: It can decide when OpenAI has reached “artificial general intelligence,” or A.G.I. That would mean OpenAI’s computers could perform most tasks that a human brain could.

Reaching A.G.I. could also reshape OpenAI’s business. When that declaration is made, Microsoft loses its rights to use OpenAI’s technology, according to the investment contract it signed with OpenAI. If OpenAI severs its ties to Microsoft, it could consider partnerships with other tech giants.

Already, OpenAI’s for-profit company has used this potential declaration as leverage against Microsoft — warning that if Microsoft will not agree to better terms, the nonprofit might issue this declaration and void their entire agreement, according to a person familiar with the company’s negotiations.

OpenAI must also satisfy another party: the public at large. In part because Mr. Altman has spent years publicly warning that A.I. could become dangerous, many individuals now share similar concerns. And many in the tech industry are publicly questioning whether OpenAI is prepared to guard against the risks its technologies will bring.

The concept of AGI is so fuzzy that such a declaration is mostly a matter of will and power. The logic will follow the party with the most power.

Sheesh! More at the link.

In the beginning: OpenAI Email Archives (from Musk v. Altman)

Habryka over at LessWrong:

As part of the court case between Elon Musk and Sam Altman, a substantial number of emails between Elon, Sam Altman, Ilya Sutskever, and Greg Brockman have been released.

I have found reading through these really valuable, and I haven't found an online source that compiles all of them in an easy to read format. So I made one.

Here's the first four emails:

Sam Altman to Elon Musk - May 25, 2015 9:10 PM

Been thinking a lot about whether it's possible to stop humanity from developing AI.

I think the answer is almost definitely not.

If it's going to happen anyway, it seems like it would be good for someone other than Google to do it first.

Any thoughts on whether it would be good for YC to start a Manhattan Project for AI? My sense is we could get many of the top ~50 to work on it, and we could structure it so that the tech belongs to the world via some sort of nonprofit but the people working on it get startup-like compensation if it works. Obviously we'd comply with/aggressively support all regulation.

Sam

Elon Musk to Sam Altman - May 25, 2015 11:09 PM

Probably worth a conversation

Sam Altman to Elon Musk - Jun 24, 2015 10:24 AM

The mission would be to create the first general AI and use it for individual empowerment—ie, the distributed version of the future that seems the safest. More generally, safety should be a first-class requirement.

I think we’d ideally start with a group of 7-10 people, and plan to expand from there. We have a nice extra building in Mountain View they can have.

I think for a governance structure, we should start with 5 people and I’d propose you, Bill Gates, Pierre Omidyar, Dustin Moskovitz, and me. The technology would be owned by the foundation and used “for the good of the world”, and in cases where it’s not obvious how that should be applied the 5 of us would decide. The researchers would have significant financial upside but it would be uncorrelated to what they build, which should eliminate some of the conflict (we’ll pay them a competitive salary and give them YC equity for the upside). We’d have an ongoing conversation about what work should be open-sourced and what shouldn’t. At some point we’d get someone to run the team, but he/she probably shouldn’t be on the governance board.

Will you be involved somehow in addition to just governance? I think that would be really helpful for getting work pointed in the right direction getting the best people to be part of it. Ideally you’d come by and talk to them about progress once a month or whatever. We generically call people involved in some limited way in YC “part-time partners” (we do that with Peter Thiel for example, though at this point he’s very involved) but we could call it whatever you want. Even if you can’t really spend time on it but can be publicly supportive, that would still probably be really helpful for recruiting.

I think the right plan with the regulation letter is to wait for this to get going and then I can just release it with a message like “now that we are doing this, I’ve been thinking a lot about what sort of constraints the world needs for safefy.” I’m happy to leave you off as a signatory. I also suspect that after it’s out more people will be willing to get behind it.

Sam

Elon Musk to Sam Altman - Jun 24, 2015 11:05 PM

Agree on all

There's much more at the link. Amazing stuff. 

Reading through the rest I'm reminded of a line from The Blues Brothers: "We're on a mission from God."